DeepStackAI's profile picture. The 1st #AI to beat pros at heads-up no-limit hold'em #poker - #DeepStackAI uses heuristic search techniques, #DeepLearning and regret minimization

DeepStack

@DeepStackAI

The 1st #AI to beat pros at heads-up no-limit hold'em #poker - #DeepStackAI uses heuristic search techniques, #DeepLearning and regret minimization

Pinned

⚡️ “DeepStack: AI beats poker pros for 1st time”


Looking to go head-to-GPU and toe-to-transistor against #DeepStackAI, the 1st #AI to beat human pros at heads-up no limit #poker? Join us tomorrow at @TWoSEdm's Dark Matters: buff.ly/2n8GnaD #Yeg


12/ Here's a summary table. The (X +/- Y) columns show estimated win probabilities, and 95% confidence, based on the W/L predictions.

DeepStackAI's tweet image. 12/ Here's a summary table. The (X +/- Y) columns show estimated win probabilities, and 95% confidence, based on the W/L predictions.

10/ Watch for an upcoming interview with @PokercastAdam and @tchanpoker on #AIVAT and these results too!


9/ Last week, we were guests on @ThinkingPoker to describe #AIVAT and its analysis of our freezeouts: thinkingpoker.net/2017/08/episod…


8/ While each freezeout always ends in a win or a loss, #AIVAT gives us win probabilities for winning a freezeout.


7/ #AIVAT also works on freezeout tournaments, like the Twitch matches we played earlier this year: youtube.com/channel/UC4vSx…


6/ And described by @IFPMatchPoker using an example hand between Martin Sturc and DeepStack: matchpoker.net/deepstack-ai/


5/ #AIVAT is provably fair, described in mathematical detail in a research paper: arxiv.org/abs/1612.06915


4/ 3,000 hands evaluated with #AIVAT is statistically equivalent to almost 100,000 hands of poker!


3/ It can then subtract these terms from the outcome, producing a "luck-adjusted" estimate with much less variance than the actual outcome.


2/ #AIVAT analyzes hands in finished matches and estimates how much of a player's winnings came from lucky or unlucky cards/actions.


1/ #DeepStackAI uses a new evaluation technique called #AIVAT for separating luck from skill in games like #poker


DeepStack reposted

Spent a really enjoyable day brainstorming AI ideas with Rich, Mike, and Patrick. So exciting to welcome them and their team to #DeepMind !

We're opening a research office in #Edmonton, Canada, in collaboration with @UAlberta, led by RL pioneer Rich Sutton deepmind.com/blog/deepmind-…

GoogleDeepMind's tweet image. We're opening a research office in #Edmonton, Canada, in collaboration with @UAlberta, led by RL pioneer Rich Sutton deepmind.com/blog/deepmind-…


DeepStack reposted

We're opening a research office in #Edmonton, Canada, in collaboration with @UAlberta, led by RL pioneer Rich Sutton deepmind.com/blog/deepmind-…

GoogleDeepMind's tweet image. We're opening a research office in #Edmonton, Canada, in collaboration with @UAlberta, led by RL pioneer Rich Sutton deepmind.com/blog/deepmind-…

DeepStack reposted

I'm heads up vs a very nice Dutchman who does not seem to have much HU experience. I, on the other hand, have trained against @DeepStackAI.

thinkingpoker's tweet image. I'm heads up vs a very nice Dutchman who does not seem to have much HU experience. I, on the other hand, have trained against @DeepStackAI.

All six of our #twitch #poker matches are now on our YouTube page: youtube.com/playlist?list=…


#DeepStackAI is proud to be teaming up with #TechAide 2017 in support of @CentraideMtl

Au cocktail de lancement #Techaide 2017 avec @mitchgarber et nos ambassadeurs engagés. #OnCentraide #gettogethertogiveback

CentraideMtl's tweet image. Au cocktail de lancement #Techaide 2017 avec @mitchgarber et nos ambassadeurs engagés. #OnCentraide #gettogethertogiveback


Don't forget to check out our YouTube account for all the latest highlights from our humans vs #AI #poker matches buff.ly/2qSJcNW


Loading...

Something went wrong.


Something went wrong.